Search code examples
docker-containertestcontainers

testcontainers-python hanging while showing "waiting to be ready...", then fails

I'm running my unit testing code for neo4j. My environment: Ubuntu 20.04LTS server 1Gb Memory 1CPU

Here is what is displayed in the console:

====================================== test session starts ======================================
platform linux -- Python 3.8.10, pytest-7.1.2, pluggy-1.0.0
rootdir: ~/morsvq, configfile: pytest.ini
plugins: mock-3.8.2
collected 2 items

---------------------------------------- live log setup -----------------------------------------
INFO     testcontainers.core.container:container.py:52 Pulling image neo4j:latest
INFO     testcontainers.core.container:container.py:63 Container started: ad7963ed01
INFO     testcontainers.core.waiting_utils:waiting_utils.py:46 Waiting to be ready...
INFO     testcontainers.core.waiting_utils:waiting_utils.py:46 Waiting to be ready...
ERROR    neo4j:__init__.py:571 Failed to read from defunct connection IPv4Address(('localhost', 49153)) (IPv4Address(('127.0.0.1', 49153)))

The same code runs successfully on a faster virtual machine with 8Gb Memory. So the code itself shouldn't be faulty. My suspision is that there is something to do with my configuration, so that it now consumes to much memory?

I've checked the official websites' documentation, but it doesn't mention the memory problem. I wonder if someone has encountered similar problem? How to fix this?

Solution

  • Disclaimer: I am a maintainer of tc-java, so I have only some basic experience with tc-python. However, some facts and constraints are universal across Testcontainers language implementations.

    As you already wrote, the code runs fine on a more powerful machine, while it fails on an extremely limited machine. 1GB of RAM is not much, I would expect it is generally not enough to successfully start a Neo4j Docker container without memory swapping. Swapping would make the startup and interactions very slow, hence the startup timeout triggers.

    For further debugging, you can run the Neo4j container directly using Docker CLI on your environment and see how it behaves.